changes related to chunks and MIDASDB

[zhaoc1]

1. Since the design of the chunks depends only on (1) the representative genomes / the pan-genome (2) chunk size. We decide to cache the chunk's design. Besides, in addition to carry the chunks design in a global variable, which can be quite heavy for large number of species of interest, we decide to only carry the cache path of the chunks for given species.
2. Add major/minor allele compute to midas_run_snps, and the user can choose to report extra columns by --advanced.
3. Clean up commands to build midas db while building it for GTDB.
4. Add midasdb_name and get_midasdb_layout to accesss to any path that is related with midas database.
5. Update unit testing codes, as well as building toy db locally.
6. Passed unit test.